IOSR Journal of Economics and Finance (IOSR-JEF)

e-ISSN: 2321-5933, p-ISSN: 2321-5925. Volume 2, Issue 2 (Nov. Dec. 2013), PP 23-26
www.iosrjournals.org

XYZ Inventory Classification & Challenges

Dinesh Kumar Dhoka1, Dr. Y. Lokeswara Choudary2
1

(Research Scholar, SRM University, Chennai, India)

(Asst. Professor, Research Guide, Dept. of Commerce, Govt. Arts College for Men, Chennai India)

Abstract : Inventory Management is a buzz word these days. As much as the industry focuses on revenue,

profits, top lines and bottom lines there is more focus these days on inventories, working capital and
efficiencies. Complexities of managing these supply chains have increased because of various factors like
globalization, increased product portfolio, decentralization etc. There are various inventory models and
matrices available today to monitor the supply chains which are mentioned in this paper but the main focus is to
study the volatility of the supply chain and its impact on Inventory.
One of the ways to classify items is based on Predictability (XYZ Analysis). Items which have
Uniform demand as X,
Varying demand as Y
Abnormal demand as Z
It is always easier to group items with similar patterns and apply appropriate strategies to each group
differently. But we need to understand if there are limitations and common underlying issues before these
strategies are employed.
Keywords; XYZ Analysis, Forecast Optimization, Inventory Planning

I.

INTRODUCTION

In any industry today forecasting and planning is a vital function. Especially when capital is required
and unnecessary inventory pile up can become a huge burden. There are un-certainties in both Supply and
Demand. Forecasting with these uncertainties does become a big challenge. It is also generally known that
forecasts have their own inherent errors [1]. With appropriate use of XYZ analysis these errors can be reduced.
The reduction of forecasting error is critical because extrapolation of historical data will amplify the inherent
errors also. As periods of planning horizon increases the forecasting errors [2] also increase.

II.

BACKGROUND TO THE STUDY

There are so many metrics that indicate the Performance of the Supply Chain [3]. The question is are
there more indicators required to understand the system better. Companys inventories are becoming more
dynamic. Not only are their product portfolios increasing, the product life cycles are reducing. This means that
companies have to handle more products and new products continuously replace the existing portfolios. This
may not seem so complicated for sales but for product design, planning and execution (production and
deliveries) they pose a lot of challenges. Is it possible to measure this dynamic characteristic of the Supply
Chain?

figure 1
The above chart Figure 1 illustrates the demand patterns of XYZ products. There is another type of
XYZ analysis that is based on the percentage of value of inventory in stock. This also follows the Pareto rule
www.iosrjournals.org

23 | Page

Xyz Inventory Classification & Challenges

according to which, 20% of the items would contribute to 80% of inventory value. But this XYZ analysis is
different from the XYZ considered in this study which is based on predictability or volatility of items.

III.

IMPORTANCE AND NEED

All the Inventory Control Methods have their limitation in terms of the usage and applicability. The
VED analysis [4] is most commonly used in spare parts management and not so popularly used in general
inventory management. FSN analysis [4] fails when used in manufacturing environment where raw materials
may be issued for production and eventually the produced items may remain in inventory giving a wrong picture
of consumption. SDE analysis mostly depends on how the vendors are managed. Strategic purchasing plays a
vital role and inaccurate information can distort the analysis. HML analysis cannot be used unless they really
have a major impact on the total inventory, in the sense that some high value items may have very low
transactions and in most cases are found in MTO and PTO situations [5]. XYZ analysis is done on inventory in
the stores which can vary dramatically every month for which the analysis is done. Various external factors like
lost/delayed sales orders and supplies can influence the analysis. The most effectively used analysis in the
industry today is ABC analysis and they have the following limitations
Highly critical parts low in consumption value may be over looked
Periodic updation and review becomes critical
Cannot precisely consider all problems of Inventory control like thousands of low value items
Money-based measures of inventory are not always best suited for operations. Also considering only
costs and compromising on other factors like service levels, efficiencies etc can impact business. Ofcourse at the
micro level customers demand at item level must be planned and controlled to match supply and demand. The
days of supply and quantities on hand may become unambiguous and does not reveal if this inventory is
appropriate.
Aggregate dollars and a historical view arent very useful to operations, which needs to match supply
and demand of specific items in future.
Various MRP tools and techniques address both of these shortcomings by coordinating the quantities
and timing of the deliveries from known target requirements to minimize cost or achieve desired service levels
[6]. With realistic lead times and realistic forecasts, the resulting inventories would be optimal for the
circumstances. Lack of realism in these areas, however, undermined the operation of MRP systems, yielding
less-than-desired service and more-than-desired inventory.

IV.

CHALLENGES IN XYZ CLASSIFICATION:

IV.1.FORMULAE AND STEPS

For XYZ analysis the following calculations have to be done.
1. Sum of Squares,
2. Variances and
3. Standard Deviation (SD).
4. Co-efficient of Variation (CV):
Compilation of all these data can be tedious and prone to errors.

=
equation 1
is Standard Deviation
Xi are individual values,
X is the average Value
N is the total number of observations
The above formula (1) may look simple for calculating SD. But if you consider data of large organization with
3000-15000 SKUs spread over a large time frame. Calculating SD can be a very tedious task. Once SD is CV is
easily calculated.
CV= / X
( is S.D. and X is Mean)
equation 2
www.iosrjournals.org

24 | Page

Xyz Inventory Classification & Challenges

The first big challenge for calculating SD.
You first need to define what your average should be?
SD needs to be calculated for weekly data or monthly data or any other data set?
Technically it is possible to estimate SD for the following periods.
1. Yearly
2. Quarterly
3. Monthly
4. Fortnightly
5. Weekly
6. Daily
7. Hourly etc
Lets say you want to calculate Variance, SD and CV for monthly data i.e. data consolidated monthly
and their monthly patterns are studied. In the example given below the table shows data consolidated for the
months Feb, Mar and Apr 2013 along with their ranks. Products against S No. 3 to 17 have the same ranks (60).
The standard deviation and the average of these 15 items are different but their CVs are the same. There is
already lot of literature that can show that both Averages and SD are not good statistics as compared to CV
especially when it comes to Inventory Management and related subjects. But in cases as shown in Table 1
below, CV can also become a huge task to handle. If 300 products have to be classified as X, Y and Z, the first
20% (X products) will have 74 products instead of 60 because of the clash in ranks
Table 1
S No

Product

Feb13

Mar13

Apr13

Std Dev

Average

CV

Rank

Prod 41

40

40

39

0.47140452

39.66666667

0.01188415

58

Prod 21

30

30

22

3.77123617

27.33333333

0.13797205

59

Prod 24

3.29983165

4.66666667

0.70710678

60

Prod 2

0.47140452

0.66666667

0.70710678

60

Prod 4

40

40

18.85618083

26.66666667

0.70710678

60

Prod 6

10

10

4.71404521

6.66666667

0.70710678

60

Prod 9

0.47140452

0.66666667

0.70710678

60

Prod 15

0.94280904

1.33333333

0.70710678

60

U
A

E
Q

Prod 20

2.82842712

4.00000000

0.70710678

60

10

Prod 25

0.94280904

1.33333333

0.70710678

60

11

Prod 3

0.47140452

0.66666667

0.70710678

60

12

Prod 21

0.47140452

0.66666667

0.70710678

60

Coeff.

13

Prod 1

0.94280904

1.33333333

0.70710678

60

Of

14

Prod 7

0.94280904

1.33333333

0.70710678

60

15

Prod 10

0.94280904

1.33333333

0.70710678

60

16

Prod 5

2.35702260

3.33333333

0.70710678

60

17

Prod 23

10

10

4.71404521

6.66666667

0.70710678

60

18

Prod 64

10

4.18993503

5.66666667

0.73940030

75

Variance

Quarterly and Yearly data would be very long periods of analysis and will have seldom usage because product
life cycles are getting shorter and shorter. Also will it be really possible to compute XYZ classification based on
Hourly, Weekly and fortnightly consolidations.
IV.2.PERIODS FOR DATA ANALYSIS
The next big challenge with XYZ classification is the total period for which the Analysis should be
done. Lets say you decide to consolidate daily data for one month. For that particular month you may have only
69 active items. If you consider longer durations the number of active items would proportionally increase as
shown in the Table 2 below.

www.iosrjournals.org

25 | Page

Xyz Inventory Classification & Challenges

Table 2
Periods

Active

1Month
2Months

69
102

3Months
4Months
5Months
6Months
7Months
8Months
9Months
12Months

133
143
164
177
182
195
203
217

% Diff

48%

13
19

30%
8%
15%
8%
3%
7%
4%
7%

25
27
31
34
35
38
39
42

% Diff

46%

23
30

32%
8%
15%
10%
3%
9%
3%
8%

42
49
52
61
64
67
71
65

% Diff

% Diff

30%

33
53

61%

40%
17%
6%
17%
5%
5%
6%
-8%

66
67
81
82
83
90
93
110

25%
2%
21%
1%
1%
8%
3%
18%

As the number of active items increase for increasing periods, proportionally the no of items classified
as Xs, Ys and Zs also increase. Now if we were to extrapolate X items for planning and budgeting purposes,
what would be the most appropriate number of Xs that we should consider. Between the two extremes 13 and 42
X Class items for 1 Month and 12 Months data analysis as shown in table above, the entire planning can go
for a toss. TO be on the safer side, instead of choosing the data on the extremes, you may settle for 5-7 Months
for data analysis for XYZ classification.

V.

DRAWBACKS FOR XYZ CLASSIFICATIONS

The most important drawback of XYZ analysis is categorization of new products. They are most often classified
as Z class items because their demand patterns are not established. The easier way is to exclude all new items
in the XYZ analysis. But if their contribution to inventory costs and sales percentages are significant, there must
be ways of quantifying this volatility created by the new items.
Another important drawback of the XYZ analysis is that the XYZ defines the predictability of the
demand among the items. There are no benchmarks or industry standards. There are no standards that say that
items which have specific variability should only be planned. Irrespective of type of industry, if X category
has significant variation in demand, it can affect overall inventory.
XYZ analysis can also overlook seasonal items. Therefore it becomes mandatory to remove these
seasonal items from this analysis. To simplify inventory management, timeframes must be defined for seasonal
products. During the season these items by default are classified as fast-moving items and after the season they
must be classified as Non-moving irrespective of the sales impact during the season.

VI.

CONCLUSION

Different Inventory Classification methods may be employed for different purposes. But we must
carefully understand their limitations and implications before major decisions are put in place based on these
classifications It would be better if some kind of automations and procedures with checks are designed for such
Analysis like XYZ where entire process may be very tedious. There may be some negative and exceptional
transactions in the database which can some impact on your XYZ data analysis. Some of the limitations of this
analysis can also be handled by introducing additional dimensions in reporting. ABC along with XYZ, a 2dimensional approach to inventory classifications can be used more effectively.

REFERENCES
Journal Papers:
[1]
[2]
[3]
[4]
[5]
[6]

Yuliang Yaoa, Rajiv Kohlib, Susan A. Sherera, Jerold Cederlundc, Learning curves in collaborative planning, forecasting, and
replenishment (CPFR) information systems: An empirical analysis from a mobile phone manufacturer, Journal of Operations
Management, Volume 31, Issue 6, September 2013, Pages 285297
Martin Stlein, John Jack Kanet, Mike Gorman, Stefan Minner, Time-phased safety stocks planning and its financial impacts:
Empirical evidence based on European econometric data, International Journal of Production Economics, April 2013
Roberto Cigolini Margherita Pero, Tommaso Rossi, Andrea Sianesi, Linking supply chain configuration to supply chain
performance: A discrete event simulation model, Simulation Modeling Practice and Theory, Volume 40, January 2014, Pages 111
Petr Besta, Kamila Janovska, Sarka Vilamova, Iveta Voznakova, Josef Kutac, The Utilisation Of Statistical Methods In The Area
Of Inventory Management, Carpathian Logistics Congress (2012)
In-Jae Jeong, A dynamic model for the optimization of decoupling point and production planning in a supply chain, International
Journal of Production Economics, Volume 131, Issue 2, June 2011, Pages 561567
Anna-Lena Beutel, Stefan Minner, Safety stock planning under causal demand forecasting, Int. J. Production Economics 140 (2012)
637645

www.iosrjournals.org

26 | Page